Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2023 May 17;13(1):8027.
doi: 10.1038/s41598-023-35227-z.

Research on tire crack detection using image deep learning method

Shih-Lin Lin  1
Affiliations

Research on tire crack detection using image deep learning method

Shih-Lin Lin. Sci Rep. .

Abstract

Driving can understand the importance of tire tread depth and air pressure, but most people are unaware of the safety risks of tire oxidation. Drivers must maintain vehicle tire quality to ensure performance, efficiency, and safety. In this study, a deep learning tire defect detection method was designed. This paper improves the traditional ShuffleNet and proposes an improved ShuffleNet method for tire image detection. The research results are compared with the five methods of GoogLeNet, traditional ShuffleNet, VGGNet, ResNet and improved ShuffleNet through tire database verification. The experiment found that the detection rate of tire debris defects was 94.7%. Tire defects can be effectively detected, which proves the robustness and effectiveness of the improved ShuffleNet, enabling drivers and tire manufacturers to save labor costs and greatly reduce tire defect detection time.

PubMed Disclaimer

Conflict of interest statement

The author declares no competing interests.

Figures

Figure 1
Figure 1
The flow chart of the traditional ShuffleNet neural network.
Figure 2
Figure 2
The flowchart of the improved ShuffleNet neural network.
Figure 3
Figure 3
Imagess of "Cracked" and "Normal" tires.
Figure 4
Figure 4
GoogleNet classification accuracy and iterations results.
Figure 5
Figure 5
GoogleNet classification loss and iterations results.
Figure 6
Figure 6
GoogLeNet confusion matrix classification results, the total correct rate is 82.7%.
Figure 7
Figure 7
Results of traditional Shufflenet classification accuracy and iterations.
Figure 8
Figure 8
Results of traditional Shufflenet classification loss and iteration times.
Figure 9
Figure 9
The traditional Shufflenet confusion matrix classification results, the total correct rate is 85.3%.
Figure 10
Figure 10
The ResNet confusion matrix classification results, the total correct rate is 90%.
Figure 11
Figure 11
The VGGNet confusion matrix classification results, the total correct rate is 87.3.
Figure 12
Figure 12
Improved Shufflenet classification accuracy and iterations results.
Figure 13
Figure 13
Improved Shufflenet classification loss and iterations results.
Figure 14
Figure 14
Improved Shufflenet confusion matrix classification results, with a total accuracy of 94.7%.
Figure 15
Figure 15
In YOLOv7, the detection results of large cracks in defective tires are 98%.
Figure 16
Figure 16
In YOLOv7, the detection results of small cracks in defective tires are 90%.
See this image and copyright information in PMC

References

    1. Zhao G, Qin S. High-precision detection of defects of tire texture through X-ray imaging based on local inverse difference moment features. Sensors. 2018;18(8):2524. doi: 10.3390/s18082524. - DOI - PMC - PubMed
    1. Guo Q, Zhang C, Liu H, Pietro S, et al. Defect detection in tire X-ray images using weighted texture dissimilarity. J. Sens. 2016;2016:1–12.
    1. Wang R, Guo Q, Lu S, et al. Tire defect detection using fully convolutional network. IEEE Access. 2019;7:43502–43510. doi: 10.1109/ACCESS.2019.2908483. - DOI
    1. Zheng Z, Zhang S, Yu B, et al. Defect inspection in tire radiographic image using concise semantic segmentation. IEEE Access. 2020;8:112674–112687. doi: 10.1109/ACCESS.2020.3003089. - DOI
    1. Li Y, et al. A novel severity calibration algorithm for defect detection by constructing maps. Inf. Sci. 2022;607:1600–1616. doi: 10.1016/j.ins.2022.06.076. - DOI